How do cardiologists select patients for dual antiplatelet therapy continuation beyond 1 year after a myocardial infarction? Insights from the EYESHOT Post‐MI Study

Abstract Background Current guidelines suggest to consider dual antiplatelet therapy (DAPT) continuation for longer than 12 months in selected patients with myocardial infarction (MI). Hypothesis We sought to assess the criteria used by cardiologists in daily practice to select patients with a history of MI eligible for DAPT continuation beyond 1 year. Methods We analyzed data from the EYESHOT Post‐MI, a prospective, observational, nationwide study aimed to evaluate the management of patients presenting to cardiologists 1 to 3 years from the last MI event. Results Out of the 1633 post‐MI patients enrolled in the study between March and December 2017, 557 (34.1%) were on DAPT at the time of enrolment, and 450 (27.6%) were prescribed DAPT after cardiologist assessment. At multivariate analyses, a percutaneous coronary intervention (PCI) with multiple stents and the presence of peripheral artery disease (PAD) resulted as independent predictors of DAPT continuation, while atrial fibrillation was the only independent predictor of DAPT interruption for patients both at the second and the third year from MI at enrolment and the time of discharge/end of the visit. Conclusions Risk scores recommended by current guidelines for guiding decisions on DAPT duration are underused and misused in clinical practice. A PCI with multiple stents and a history of PAD resulted as the clinical variables more frequently associated with DAPT continuation beyond 1 year from the index MI.


| INTRODUCTION
Current guidelines suggest that continuation of dual antiplatelet therapy (DAPT) for longer than 12 months should be considered in patients with myocardial infarction (MI) who have tolerated DAPT without bleeding complications. 1,2 This recommendation is based on recent randomized clinical trials suggesting that DAPT prolongation reduces the rate of recurrent ischemic events in post-MI patients, compared with aspirin alone. 3,4 However, these benefits come at the cost of increased risk of bleeding, raising the question about how to identify the ideal patient profile who could safely prolong DAPT. 1,2,5,6 Although predictors of DAPT interruption have been already identified in large international registries, [7][8][9] specific appraisals underlying the decision on the extension of DAPT have been poorly investigated. 10,11 Using data from the EYESHOT (EmploYEd antithrombotic therapies in patients with acute coronary Syndromes HOspitalized in iTaly) Post-MI study 12 we sought to evaluate the criteria used by cardiologists in daily clinical practice for selecting post-MI patients eligible for DAPT continuation beyond 1 year.

| METHODS
The methods used for the EYESHOT Post-MI study have been described previously. 12 Briefly, it was a prospective, observational, nationwide registry of consecutive patients with a prior MI managed by cardiologists. All patients admitted in cardiology units and/or ambulatory clinics during a period of 3 months with a documented history of presumed spontaneous MI event (non-ST-elevation, NSTEMI or ST-elevation-MI, STEMI) occurred between 1 and 3 years before enrolment have been included. 12 Patients were enrolled at the beginning of outpatient or day-hospital visit or at hospital admission.
Physicians were asked to report medications at enrolment and at the end of the visit or hospital discharge.  12 The study has been carried out in 165 cardiology centers. Most of the patients were enrolled during outpatient or day-hospital visits (84%) and the remaining during hospital admissions (16%). 12 Data were collected using a web-based, electronic case report form with the central database located at the ANMCO Research Center. By using a validation plan, integrated in the data entry software, data were checked for missing or contradictory entries and values out of the normal range.

| Statistical analysis
Categorical variables are presented as number and percentages and compared by the χ 2 test. Normally distributed, continuous variables are presented as mean and SD (SD) and compared by means of t test, while not normally distributed variables are reported as median and interquartile range (IQR)and assessed by the Mann-Whitney U test.
The study cohort was stratified according to the treatment with DAPT at the time of enrolment.
Clinically relevant variables were included in a multivariable model (logistic regression), to identify the independent predictors of DAPT assumption at the time of enrolment and after hospital discharge from a cardiology ward or at the end of the visit by cardiologist among patients in the second and the third year from the index MI. The variables included in the logistic model were: age, gender, diabetes mellitus, renal insufficiency, peripheral artery disease (PAD), history of major bleeding events or surgery, heart failure, atrial fibrillation, presence of symptoms, prior revascularization, number of stent implanted (≤2 vs >2), type of last myocardial infarction (STEMI vs NSTEMI), type of enrolment (hospital admission vs outpatient visit). A P value < .05 was considered statistically significant. All tests were two-sided. Analyses were performed with SPSS system software, version 24.

Out
Baseline characteristics of patients on DAPT vs those not receiving DAPT at enrolment are shown in Table 1. Patients on DAPT were on average younger (64 vs 66 years, P = .002) and had a shorter time since index MI (20 vs 23 months, P < .001). In addition, patients on DAPT had less frequently atrial fibrillation (6.8% vs 15.9%, P < .0001) but did have a higher incidence of PAD compared to patients without DAPT at enrolment (9.3% vs 5.6%, P = .005).
Among patients receiving DAPT, the most frequently used combination was aspirin and clopidogrel for both patients in the second and third year from MI ( Table 2).
Scores for the assessment of ischemic or bleeding risk were used in 8.1% and 6.9% of patients, respectively ( Figure 2). The GRACE and the HAS-BLEED resulted as the risk scores mostly used for the evaluation of ischemic or bleeding risk, respectively ( Figure 2). Ischemic risk scores were used in 9.5% of the patients on DAPT and in 7.4% of patients without DAPT (P = .153); whereas bleeding risk scores were used in 8.1% and 6.2% of patients on DAPT and without DAPT, respectively (P = .179).
At multivariate analyses, PCI with multiple stents and PAD resulted as independent predictors of DAPT continuation, while atrial fibrillation was the sole independent predictor of DAPT interruption for both patients at the second and the third year from MI at enrolment and the time of discharge/end of the visit (Figure 3).

| DISCUSSION
The present analysis of a nationwide study on consecutive patients managed by cardiologists 1-3 years after a MI demonstrates that: In post-MI patients, the cardiovascular risk remains substantially elevated beyond the first year. [13][14][15] The REACH (REduction of Atherothrombosis for Continued Health) registry showed an incidence of recurrent cardiovascular events of 18% at 4 years in patients with history of MI or stroke. 13 Accordingly, the APOLLO dataset, which links registries and administrative data, demonstrated that the risk of cardiovascular events in MI survivors is approximately 20% across the first 3 years from MI. 14,15 In this setting, prolonged DAPT has been shown to be an effective therapeutic strategy to prevent recurrent ischemic events. [1][2][3][4] Nonetheless, because continued antiplatelet therapy is also associated with increased risk of bleeding, it is necessary to weigh this risk against the potential benefit. [1][2][3][4] Decisions about duration of DAPT are best made on an individual basis and should integrate several clinical variables. In this regard, novel risk scores have been specifically designed to guide and inform decision making for optimal DAPT duration. 16 However, most of the frequently used risk scores for assessing ischemic events or major bleedings in our cohort were originally developed and validated for the prediction of events occurring mainly during hospital stay or at short term follow-up after a MI or in the setting of atrial fibrillation. 17 F I G U R E 2 Use of scores for the assessment of ischemic or bleeding risk implantation and subsEquent Dual Anti Platelet Therapy) scores, 19,20 have been extremely underused in our registry.
Several studies demonstrated major bleeding events and urgent surgery are the most common reasons for early DAPT interruption. [7][8][9] However, few have investigated the reasons why cardiologists prolong or interrupt DAPT beyond the first year after a MI. 10,11 The EPI- On the other hand, the presence of atrial fibrillation needing oral anticoagulation therapy was an independent predictor of DAPT interruption in our analysis. This finding is in accordance with recent guidelines and consensus documents recommending a shorter duration of DAPT in order to reduce the risk of major bleeding events. 1,2,24

| Study limitations
Our study must be evaluated in the light of the known limitations of observational, cross-sectional studies. In addition, even if the participating centers were asked to include in the registry all consecutive post-MI patients, we were not able to verify the enrolment process, due to the absence of administrative auditing. We believe that it is unlikely however that selective enrolment in few sites may have sub-

ACKNOWLEDGMENTS
The authors would like to thank Barbara Bartolomei Mecatti for editorial assistance.
The sponsor of the study was the Heart Care Foundation, a nonprofit independent organization, which also owns the database. Database management, quality control of the data and data analyses were under the responsibility of the ANMCO Research Centre Heart Care Foundation. The study was partially supported by an unrestricted grant by Astra Zeneca, Italy. No compensation was provided to participating sites, investigators, nor members of the Steering Committee.
The Steering Committee of the study had full access to all of the data in this study and takes complete responsibility for the integrity of the data and the accuracy of the data analysis.

CONFLICT OF INTEREST
L.D.L. reports personal fees from Astra Zeneca and Daiichi Sankyo outside the submitted work. All other authors have reported that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.